Sensitivity Analysis of the Affecting Parameters on the Urban Runoff Results in the SWMM Model (Case Study: North Catchments of Tehran City)

Awareness of the sensitivity of the input data of simulation models is important for their development and application, and leads to better understanding and better estimation of values and reduces uncertainties. In this study SWMM software has been calibrated with real meteorological and hydrometric data for the part of the northern catchments of Tehran, simulation parameters have been obtained. For this purpose, five rainfall events and runoff data related to these rainfalls, recorded at the outlet of Zargandeh catchment were used. This model is calibrated with three events and verified with two other events. By comparing the results of observational and computational, the output flows of Zargandeh basin, the root mean square error (RMSE) are obtained for the first to fifth events were 0.05, 0.22, 0.39, 0.37 and 0.16, and the Nash-Sutcliffe coefficient (NS) are obtained 0.91, 0.94, 0.93, 0.9 and 0.94, respectively, that shows the appropriate adaptation of the model output results with the real observational values. The aim of this study was to investigate the effect of sensitivity of seven input parameters of SWMM model for the north catchments of Tehran city. Based on comparing the Sensitivity Index (I), Impervious area percentage (Imp%), N-Manning of impermeable areas (Nimp), Width equivalent (W) and slope (S0) of sub-basins had the greatest impact on the output results and  the depth of depression storage (Dp) and N-manning of permeable areas (Np) were diagnosed as the least effective input parameters. The effect of increasing three parameters, percentage of impermeable lands (Imp%), equivalent width (W) and slope (S0) of subbasins on the output results of the model is incremental and the reaction of the output results to increasing other parameters such as Manning roughness coefficient (Nimp) and storage depth of impermeable (Dp) areas is decreasing. Also, the sensitivity index (I) of peak flood flow, runoff volume and runoff depth for the percentage of impermeable lands (Imp%) and the width equivalent (W) of sub-basins are higher in mountainous areas than urban areas of this study area. The reason can be considered as a higher percentage of mountainous areas (84%) than urban areas (16%) in this study. On the other hand , the results showed that 10% increasing of the impermeable areas (Imp%), 5% peak runoff flow rate, 7.79% runoff volume and 5% runoff depth increase Also, 10% decreasing of the roughness coefficient of impermeable areas (Nimp), 4% of maximum runoff flow, 1.7% of runoff volume and 3% of runoff depth at the outlet point of the catchment area increased. However, land use changing and urbanization, which results of increasing the impermeability of the urban surfaces and reducing the roughness coefficient of these areas, can increase the peak flow of floods, the volume of runoff and the overflow of runoff from urban canals and it causes of the property and human losses risk. The results of this study can be useful for future modeling in the northern part of Tehran city and other areas that are similar to the catchment in terms of the hydraulic and physiological characteristics..